Full Terms & Conditions of access and use can be found athttp://www.tandfonline.com/action/journalInformation?journalCode=tjoo20

Download by: [Alma Mater Studiorum - Università di Bologna] Date: 07 February 2017, At: 07:17

Journal of Operational Oceanography

ISSN: 1755-876X (Print) 1755-8778 (Online) Journal homepage: http://www.tandfonline.com/loi/tjoo20

Operational oceanography for the Marine StrategyFramework Directive: the case of the mixingindicator

C. Fratianni, N. Pinardi, F. Lalli, S. Simoncelli, G. Coppini, V. Pesarino, A.Bruschi, M.L. Cassese & M. Drudi

To cite this article: C. Fratianni, N. Pinardi, F. Lalli, S. Simoncelli, G. Coppini, V. Pesarino, A.Bruschi, M.L. Cassese & M. Drudi (2016) Operational oceanography for the Marine StrategyFramework Directive: the case of the mixing indicator, Journal of Operational Oceanography,9:sup1, s223-s233, DOI: 10.1080/1755876X.2015.1115634

To link to this article: http://dx.doi.org/10.1080/1755876X.2015.1115634

Published online: 10 Jun 2016.

Submit your article to this journal

Article views: 36

View related articles

View Crossmark data

Operational oceanography for the Marine Strategy Framework Directive: the caseof the mixing indicatorC. Fratiannia, N. Pinardia,b, F. Lallic, S. Simoncellia, G. Coppinid, V. Pesarinoc, A. Bruschic, M.L. Cassesec andM. Drudia

aIstituto Nazionale di Geofisica e Vulcanologia, Sezione di Bologna, Italy; bDepartment of Physics and Astronomy, Alma Mater Studiorum,University of Bologna, Italy; cItalian Institute for Environmental Protection and Research, Rome, Italy; dCentro EuroMediterraneo suiCambiamenti Climatici, Lecce, Italy

ABSTRACTIn this paper we show that operational oceanography products can be used to develop indicators asrequired by the Marine Strategy Framework Directive (MSFD). We present a mixing indicator that iscalculated using MyOcean Marine Service reanalysis products. Seasonal climatology data of theBrunt–Väisäla frequency (BVF) were computed for 2001–2010 and the vertical mixing coefficientwas defined. A vertical mixing indicator was then computed in order to differentiate betweendifferent mixing conditions depending on the seasons and differentiating between the shelf andthe open ocean in the central Mediterranean Sea.

Introduction

The Mediterranean marine ecosystem is subject to astrong pressure related to pollution, coastal develop-ment, climate change and overfishing. Marine environ-ment protection, preservation and restoration, whenpossible, requires an integrated approach together withthe continuous monitoring and modelling of physicaland biochemical processes. The Marine Strategy Frame-work Directive (MSFD) embraces such an approach andrequires European member states to prepare nationalplans for monitoring and modelling the state of the marineecosystem, from the coasts to offshore areas (Figure 1).MSFD reporting should also be coordinated in accordancewith theWater Framework Directive (WFD) and the Med-iterranean Pollution Monitoring and Research (MEDPOL)programme of the Barcelona Convention. Meiner (2010)report that future developments in MSFD assessmentsshould contribute to a better characterisation of maritimespace and marine ecosystems, hand in hand with theCopernicus Marine Environment Service (CMES) whichproduces space–time-consistent data for the developmentof geospatial information systems for the integrated man-agement of the marine territory.

Operational oceanography scientific approach andpractices are the basis of the CMES, which was firstimplemented in the Mediterranean Sea in 2000(Pinardi et al. 2003; Pinardi & Coppini 2010). Today,the Copernicus programme – previously known as the

Global Monitoring for Environment and Security(GMES) programme – provides accurate, timely andeasily accessible information to improve the manage-ment of the terrestrial, marine and atmospheric environ-ment. Since 2009, a pre-operational marine service hasbeen in place with the MyOcean projects (Bahurelet al. 2009) delivering products on the ocean state fromthe global to regional scales (main European basinsand seas).

The MyOcean Marine Service at present deliversvalue-added information, including observations, ana-lyses, reanalyses and forecasts describing the physicalstate of the ocean and its primary biogeochemical par-ameters. It also contributes to research on the climateby providing long time series of reanalysed parameters.

The Mediterranean Forecasting System (MFS) is cur-rently one of the monitoring and forecasting productioncentres of the pre-operational MyOcean service (Dom-browsky et al. 2009). The system produces 10-day fore-casts daily for several ocean state variables such ascurrents, temperature, salinity and sea level. The systemalso provides a so-called reanalysis – an optimal recon-struction of the state of the marine environmental vari-ables for the Mediterranean Sea over the past fewdecades (Adani et al. 2011).

The MSFD defines 11 descriptors of good environ-mental status (GES). This paper focuses on No. 7,‘hydrographic conditions’ and the physical featurescontained in Table 1 of Annex III of the MSFD

© 2016 Institute of Marine Engineering, Science & Technology

CONTACT C. Fratianni claudia.fratianni@ingv.it

JOURNAL OF OPERATIONAL OCEANOGRAPHY, 2016VOL. 9, NO. S1, s223–s233http://dx.doi.org/10.1080/1755876X.2015.1115634

(European Union 2008). Among other characteristicsrequired, ‘mixing’ is listed as follows: in oceanographymixing regulates the stratification/destratification cycleof the water column from seasonal to hourly scales(Gaspar et al. 1990; Moum & Smyth 2001). This is con-nected to the oxygenation of the water column whichoccurs through air–sea gas exchanges and is a primary indi-cator of GES in the shelf areas of the European seas. Lowmixing allows eutrophication conditions to produce oxygendepletion with a subsequent impact on the ecosystem’shealth and productivity. Rinaldi et al. (this issue) addressseveral additional MSFD-relevant indicators derived fromin situ and satellite-based observations and made routinelyavailable within operational marine services.

In this paper we construct a mixing indicator from theMyOcean Marine Service products at the MediterraneanSea basin scale. Firstly, the Brunt–Väisäla frequency(BVF) is calculated from the reanalysis temperatureand salinity fields for the shelf and open ocean areas ofthe Central Mediterranean Sea. Secondly, the BVF isconnected to a vertical mixing coefficient. After athreshold has been set between low and high mixing con-ditions, a mixing indicator is defined for each

climatological month and for the Central MediterraneanSea areas, covering the relevant MSFD regions.

The paper is organised as follows. In the next sectionwe describe the MSFD Italian sea implementation areas,which is followed by a description of the reanalysis dataand the method used to compute the BVF. The mixingindicator is then defined in and some conclusions aredrawn in the final section.

TheMSFD and the Italian sea assessment areas

The activities which unfold in the marine and maritimesectors have always been of crucial importance to thesocio-economic and cultural development of Europe.The European seas and coasts are a vital resource, pro-viding Europeans with natural resources as well as accessto trade and transport. In general all activities related toit are very important components of the European grossnational product (GNP) – and the sea, in particular theMediterranean, is a challenging natural and cultural heri-tage. Over the last decade, the awareness of the need foran integrated approach for the protection, preservationand restoration of the marine environment has grown.

Figure 1. The various marine territorial waters as defined by the United Nations Convention of the Law of the Sea (UNCLOS) with theindication of the area that is considered by the international agreements (MEDPOL) and the European directives (WFD and MSFD) con-sidered in this paper.

Table 1. Legislative references for the various MSFD implementation stages.Actions Article

Initial assessment of the current environmental status of national marine waters and an analysis of theenvironmental and socio-economic impact of human activities on these waters.

Art. 8 – Annex III

Determination of a set of characteristics for GES for the marine waters. Art. 9 – Annex IEstablishment of a comprehensive set of environmental targets and associated indicators to guide progresstowards achieving GES in marine waters.

Art. 10 – Annex IIIand IV

Establishment and implementation of monitoring programmes for the ongoing assessment of theenvironmental status of the marine waters and the regular update of targets.

Art. 11 – Annex IIIand V

Establishment and practical implementation of measures to achieve the targets. Art. 13

s224 C. FRATIANNI ET AL.

The pillar for this integrated policy is Directive 2008/56/EC, which establishes a framework for communityaction in the field of the marine environmental policy,known as MSFD, formally adopted by the EuropeanUnion (EU) in July 2008 and transposed into Italianlaw in 2010 (Decree No. 190) (European Union 2008).

The directive is the basis of the Integrated EuropeanMaritime Policy and sets out a common frameworkbased on cooperation between member states to promotethe sustainable use of the seas and the conservation ofmarine ecosystems. The directive establishes a frame-work within which member states must take the necess-ary measures to achieve or maintain GES of their marinewaters by 2020, thus confirming the fundamental roleplayed by environmental monitoring in such an assess-ment. Each member state is required to develop a marinestrategy for its marine region or subregion, subdividedinto a preparation phase and a programme of measures,all of which is described in various articles of the direc-tive (Table 1).

Each member state is required to report on the currentstate of nature and environment of their marine waters (inaccordance with Art. 8) and define the desired state oftheir marine waters (Art. 9). In case of a deviation betweenthe current and desired state of marine waters, in order toachieve or maintain GES, the member state must establishenvironmental targets (Art. 10). These first three stagesmust be reported to the European Commission (EC) by15 October 2012. Based on these reports, the EC has to

‘assess whether, in the case of each Member State, theelements notified constitute an appropriate frameworkto meet the requirements of this Directive (Art.12)’ (Euro-pean Union 2008). By 2015, the Programme of Measures(Art. 13) must be established, based upon the first threestages (initial assessment, GES and targets) and the prep-aration of monitoring programmes (Art. 13). MSFDimplementation is a step-by-step process in which eachstep builds upon the previous one. Since the directive fol-lows an adaptive management approach, these stagesshould be kept up to date and reviewed every six yearsto ensure that the aims of the directive are adequatelymet, adjusting GES targets, monitoring and programmesof measures where necessary (Art. 17) (European Union2008).

The MSFD applies to the area of marine waters overwhich a member state exercises jurisdictional rights(Figure 1). The Mediterranean Sea is subdivided intofive marine subregions: the Western MediterraneanSea, the Adriatic sea, the Ionian Sea, the Central Medi-terranean Sea and the Aegean-Levantine Sea (Figures2–3).

Member States can define specific subdivisions withintheir (sub)region based on the specific nature of that areaand conduct separate assessments for each subdivision.For this work, the Italian seas were subdivided intoeight assessment areas (Figure 4): the Northern AdriaticSea (1) , Central Adriatic Sea (2), Southern Adriatic Sea(3), Ionian Sea (4), Central Mediterranean Sea (5), Tyr-rhenian Sea (6), Ligurian Sea (7) and Sardinian Sea (8).This subdivision is mainly geomorphological, consideringdeep basins and straits separately.

Data and methods

The data used in this work is the reanalysis from theMyOcean Marine Service (www.myocean.eu). The rea-nalysis covers the 10-year period 2001–2010 and con-siders all the available in situ temperature and salinityobservations, along-track satellite sea level anomalies(SLAs) and satellite sea surface temperatures (SSTs) toproduce gridded optimal estimates of the three-dimen-sional thermohaline, sea level and current conditions inthe Mediterranean Sea. The grid resolution is 1/16 deg(∼6.5 km) in the horizontal and 71 unevenly spacedlevels in the vertical. This reanalysis was produced forthe MyOcean Marine Service following the same schemepresented in Adani et al. (2011) but only real-time datawere used since most of the in situ temperature and sal-inity data in this period came from the voluntary observ-ing ship (VOS) programme for the Mediterranean Sea(Manzella et al. 2007) and the Argo profiling float pro-gramme (Poulain et al. 2007).Figure 2. Implementation stages for the MSFD.

JOURNAL OF OPERATIONAL OCEANOGRAPHY s225

The relationship between vertical mixing andstratification

Vertical mixing in the ocean is a key driver of the marineecosystem response, from the coastal areas to the openocean. The marine food web (Pinardi et al. 2006) devel-ops differently depending on the marine stratificationconditions, which are themselves connected to differentvertical mixing processes. In the Mediterranean Seacoastal areas, mixing occurs in winter and produces avery homogeneous water column, which in turn is domi-nated by a herbivorous food web and highly oxygenatedwaters. On the other hand, during summer both in thecoastal and open ocean areas, stratification may be

persistent, which can increase the carbon flux throughthe bacteria components of the food web, thus producingoxygen consumption and possible hypoxia, potentiallyeven resulting in anoxic conditions.

Vertical mixing is dependent on the momentum andbuoyancy fluxes at the air–sea interface and the pre-existing water column vertical stratification as well asthe energy dissipation rate. Vertical stratification isimportant in the mixing process since more turbulentkinetic energy from the wind is required to displacewater across a strong vertical density gradient (Gargett& Holloway 1984). In the open waters of the Mediterra-nean Sea, the vertical stratification undergoes seasonalchanges (Hecht et al. 1988; Lascaratos 1993; Artegianiet al. 1997a, 1997b; Pinardi et al. 2015) while near thecoasts it is maintained by the river runoff. During thewinter, the Mediterranean Sea open ocean and coastalwaters become vertically homogeneous due to theintense air–sea interaction fluxes which create favourableconditions for high mixing. In some coastal areas, on theother hand, where the river input is persistent, stratifica-tion is permanent.

The stratification can be quantified by the buoyancyfrequency or BVF, which is defined as

N =���������− gr0

drdz

√(1)

where g is the gravity acceleration (9.8 m s−2), ρ0 is themean density and dr/dz is the local vertical density gra-dient. The depth of the BVF maximum is normally takento be representative of the pycnocline depth.

According to Osborn (1980), the relationship betweenthe BVF and vertical mixing coefficient, K, or vertical

Figure 3. Subregions for Mediterranean Sea according to MSFD.

Figure 4. Italian MSFD assessment subregions.

s226 C. FRATIANNI ET AL.

diffusivity, can be expressed as

K = 1GN−2 (m2s−1) (2)

where the mixing efficiency G is assumed to be equalto 0.2 following Shih et al. (2005) and the rate of turbu-lent kinetic energy dissipation for the Mediterranean Seais set as 1 = 1.510−9m2s−3 as an average from the valuesof Cuypers et al. (2012). From Equation (2) it is evidentthat high and low values of N2, indicating high and lowstratification, respectively, are inversely correlated to thevertical mixing – therefore, calculating the BVF enablesK to be defined as a mixing indicator.

Analysis of the climatological stratification

Using the temperature and salinity profiles from the rea-nalysis data set, the average BVF was estimated in each ofthe eight assessment areas (Figure 4) from the surface to200 m. The squared BVF seasonal profiles (consideringthree-month periods starting from January) are plottedin Figure 5. The Adriatic Sea (regions 1, 2 and 3 inFigure 5) is characterised by the largest squared BVFvalues of the different areas, and the maximum valuesare reached in the first 80–100 m. In the other areas,the maximum values are found in the first 140 m whilethe squared BVF is practically zero below such depths.The squared BVF vertical profiles show an ample seaso-nal stratification cycle, from the maximum valuesattained in the homogenisation period in the first 20 mduring summer and the minimum values reached duringwinter. The BVF value distinguishing between autumn/winter and spring/summer conditions in Figure 5 is0.1 10−3 s−2 as indicated by the vertical grey line. Thusthis was chosen as the threshold N2

0 to define well-stratified versus non-stratified water column conditions.

In order to estimate a vertical average stratificationvalue for each assessment area, the average volumesquared BVF values were taken, where the volume isdefined as the assessment area multiplied by the watercolumn depth up to 100 m. The average volume squaredBVF values for each region are shown in Figure 6. It isnow evident that region 1 (the Northern Adriatic area)is always stratified on the Italian side (since it is influ-enced by the Po River), while all the other areas startto be stratified in April/May and become homogeneousfrom October to March.

Figure 7 shows the seasonal horizontal distribution ofthe vertical mixing coefficient calculated using Equation(2), averaged along the first 100 m for all the assessmentareas. As expected, the values vary both in time and spacebetween open ocean and coastal areas: areas of high ver-tical mixing develop all along the coastlines of the basin

during winter, while the open ocean shows relatively lowvalues all year long.

The mixing indicator

In order to define an indicator, we need to set thresholdsand reference values for K so that we can then obtain anindicator referring to occurrences of higher and lowervalues than the threshold for the quantity of interest.As shown in Figure 7 the vertical mixing conditionsfor the Italian seas are very variable and distinct betweenthe open ocean and coastal zones. If we want to get agood indicator of vertical mixing, the latter should beable to distinguish between these regions for each area.

Using Equation (2) we define a threshold mixing indi-cator based on the N2

0 values established in Figures 5 and6 thus differentiating between stratified and non-strati-fied conditions during summer and winter. Conse-quently the vertical mixing threshold value is

K0 = N−20 1G (3)

Invoking the values for 1, G, N20 indicated above, the

threshold value of the mixing parameter is K0 = 310−6

m2 s−1. This value is one order of magnitude largerthan molecular thermal diffusivity (approx. 10−7 m2

s−1) and accounts for the turbulent mixing of theupper water column. The mixing indicator, Imx, isthen defined as

Imx(xi, yi, tj) = 100

∑Nk=1 P(xi, yi, t jk)

N, (4)

where P is set to

P = 1 if K . K0,

P = −1 if K , K0.(5)

The sum in Equation (4) is intended for each availableyear k of each month j, which in our case provides 10different realisations for each month from 2001 to2010. In addition,(xi, yi) indicate the model field gridpoints. In summary, low and high mixing coincidewith negative and positive values, respectively.

The resulting monthly maps of Imx are displayed inFigure 8. All the areas show high values of the mixingindicator for January to April with the exception of theNorthern Adriatic Sea and the Po-River-influenced mar-ine waters which always have low mixing indicatorvalues. In the transition season (May/June) and summer(July–September) all the assessment areas show low mix-ing values from the open ocean to the coastal areas. Theonly exception is the extended area of the Tunisian shelfin the Strait of Sicily where relatively high mixing con-ditions are found all year long. The mixing indicator is

JOURNAL OF OPERATIONAL OCEANOGRAPHY s227

Figure 5. Averaged area seasonal squared BVF profiles for each Italian assessment area. Note: The vertical grey line corresponds to theN20 threshold (10–4 s–2).

s228 C. FRATIANNI ET AL.

also relatively low in October in the Southern Ionian Sea(region 4 in Figure 4). This is a region well known for itsanticyclonic circulation (Pinardi et al. 2015), whichmaintains the summer stratification longer than inareas of opposite sign circulation. The Ligurian Sea, theNorthern Tyrrhenian Sea (regions 7 and 6 in Figure 4,

respectively) and the Northern Adriatic Sea (region 1in Figure 4) show an increase in the mixing rate in Octo-ber due to the favourable pre-conditioning cyclonic cir-culation in these areas.

There is a narrow band of high mixing values all alongthe Croatian and Italian coastlines in September and

Figure 6. Averaged volume squared BVF monthly climatology for each of the assessment areas in Figure 4.

JOURNAL OF OPERATIONAL OCEANOGRAPHY s229

Figure 7. Monthly horizontal distribution of the first 100 m averaged vertical mixing coefficient calculated from Equation (2).

s230 C. FRATIANNI ET AL.

Figure 8. Monthly maps of the mixing indicator for all the Italian assessment areas.

JOURNAL OF OPERATIONAL OCEANOGRAPHY s231

October, indicating that these shelf areas already havedetectable and specific vertical mixing conditions withrespect to the adjacent open ocean areas.

Several regions show relatively large mixing indicatorgradients throughout the year. In August and September,the Northern Adriatic Sea (region 1 in Figure 4) showscontrasting mixing values between the Gulf of Trieste,the Northern Adriatic Sea and the Po-River-influencedwaters where the minimum values of mixing are found.In May the southern regions of the Strait of Sicily and theIonian Sea (regions 4 and 5 in Figure 4) show relativelylarge values of the mixing indicator, while all otherregions already show minimum values.

Conclusions

We have presented an application of the MyOcean Mar-ine Service products which is used to define a mixingindicator for the MSFD assessment of ‘hydrographicconditions’. Among all MyOcean Marine Service pro-ducts, here we use the reanalysis and optimal reconstruc-tion of the thermodynamic characteristics of theMediterranean Sea from observations and a numericalmodel. The high space–time coverage of the reanalysisproducts resolves the seasonal cycle and the major differ-ences between the shelf and open ocean areas.

The vertical mixing coefficient is defined as beinginversely proportional to the BVF squared, which canbe computed from the reanalysis temperature and sal-inity fields. The mixing indicator is then defined as thenumber of occurrences where the value of the mixingcoefficient is larger than a given threshold. The mixingcoefficient threshold only depends on the thresholdvalue of the BVF. A comprehensive subregional analysisof the BVF seasonal cycle was thus carried out in order toset the threshold. A threshold value of N2

0 = 10−4 allowsdifferentiation between stratification and destratificationseasons and consequently the threshold value of mixingwas determined to be K0 = 310−6 m2 s−1.

The low mixing conditions permanently found in thePo River have been shown to influence waters thatextend several hundred km from the river mouth, high-lighting the area’s particular sensitivity to oxygen con-sumption even during the winter months. Most of theother shelf areas, and in particular the Tunisian shelf(the southern part of region 5 in Figure 4), are generallycharacterised by high vertical mixing, although themodel resolution cannot fully resolve some of the narrowshelf areas prevailing in most of the investigated subre-gions. Particular evidences of high mixing conditionsalong the shelves include the Eastern Adriatic Sea, theNorthern Tyrrhenian Sea and the Ligurian Sea (regions6 and 7 in Figure 4).

The open ocean areas were found to be affected by highmixing during the winter and lowmixing during the sum-mer. This is a specific characteristic of the MediterraneanSea due to the high seasonality of the winds and the highstratification conditions encountered during the summerdue to the rather intense warming of the sea surface atthese latitudes. In general though, the basin completelyloses its summer thermocline and does not develop amain deep thermocline (Pinardi & Masetti 2000).

The mixing indicator developed here, based on theMyOcean Marine Service Mediterranean Sea reanalysis,could contribute to the integrated assessment of GES.Although still in its early stages, the method seems tobe robust. Future research could consider a longer timeseries of reanalysis data and different values of the BVFthreshold depending on the Central Mediterranean Seasubregions.

Acknowledgements

This work was carried out within the framework of the scien-tific collaboration agreement for the Marine Strategy betweenthe ISPRA and the INGV.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

The work was supported by the MyOcean2 project [grantnumber 283367]; Nextdata: National Research Programme(PNR) 2011-2013-Founded by the Italian Ministry of Univer-sity and Research (MIUR); and RITMARE: National ResearchProgramme (PNR) 2012-2016-Founded by the Italian Minis-try of University and Research (MIUR).

References

Adani M, Dobricic S, Pinardi N. 2011. Quality assessment of a1985–2007 Mediterranean sea reanalysis. J Atmos OceanicTechnol. 28:569–589. doi:10.1175/2010JTECHO798.1

Artegiani A, Bregant D, Paschini E, Pinardi N, Raicich F,Russo A. 1997a. The Adriatic sea general circulation. PartI: air-sea interactions and water mass structure. J PhysOceanogr. 27:1492–1514. doi:10.1175/1520-0485(1997)027<1492:TASGCP>2.0.CO;2

Artegiani A, Bregant D, Paschini E, Pinardi N, Raicich F,Russo A. 1997b. The Adriatic sea general circulation. PartII: baroclinic circulation structure. J Phys Oceanogr.27:1515–1532. doi:10.1175/1520-0485(1997)027<1515:TASGCP>2.0.CO;2

Bahurel P, Adragna F, Bell MJ, Jacq F, Johannessen JA, LeTraon P-Y, Pinardi N, She J. 2009. Ocean monitoring andforecasting core services: the European MyOceanExample. Paper presented at OceanObs ‘09 Ocean

s232 C. FRATIANNI ET AL.

Information for Society: Sustaining the Benefits, Realizingthe Potential, September 21–25, Venice, Italy.

Cuypers Y, Bouruet-Aubertot P, Marec C, Fuda J-L. 2012.Characterization of turbolence from a fine-scale parametriza-tion and microstructure measurements in the Mediterraneansea during the BOUM experiment. Biosciences. 9:3131–3149.doi:10.5194/bg-9–3131–2012

Dombrowsky E, Bertino L, Brassington GB, Chassignet EP,Davidson F, Hurlburt HE, Kamachi M, Lee T, Martin MJ,Mei S, Tonani M. 2009. GODAE systems in operation.Oceanography Magazine. 22:0–95.

European Union. 2008. Directive 2008/56/EC OJ L 164 of25.6.2008. 19–40.

Gargett AE, Holloway G. 1984. Dissipation and diffusion byinternal wave breaking. J. Mar. Res. 42:15–27.

Gaspar P, Grégoris Y, Lefrevre J-M. 1990. A simple eddy kin-etic energy model for simulations of the ocean vertical mix-ing: test at station Papa and long-term upper ocean studysite. J Geophys Res. 95:16179–16193.

Hecht A, Pinardi N, Robinson AR. 1988. Currents, watermasses, eddies and jets in the Mediterranean LevantineBasin. J Phys Oceanogr. 18:1320–1353.

Lascaratos A. 1993. Estimation of deep and intermediate watermass formation rates in the Mediterranean sea. Deep SeaRes II. 40:1327–1332.

Manzella GMR, Reseghetti F, Coppini G, Borghini M, CrusadoA, Galli C, Gertman I, Hayes D, Millot C, Ozsoy E, et al.2007. The improvements of the ships of opportunity pro-gram in MFSTEP. Ocean Sci. 3:245–258. www.ocean-sci.net/3/245/2007/

Meiner A. 2010. Integrated maritime policy for the EuropeanUnion consolidating coastal and marine information tosupport maritime spatial planning. Journal of CoastConserv. 14:1–11. doi:10.1007/s11852–009–0077–4

Moum JN, SmythWD. 2001. Upper ocean mixing. Encyclopediaof Ocean Sciences. 6 Academic Press: 3093–3100.

Osborn TR. 1980. Esitmates of the local rate of vertical diffusionfrom dissipation measurements. J Phys Oceanogr. 10:83–89.

Pinardi N, Allen I, Demirov E, De Mey P, Korres G, LascaratosA, Le Traon P-Y, Maillard C, Manzella G, Tziavos C. 2003.The Mediterranean ocean forecasting system: first phase ofimplementation (1998–2001). Ann Geophys. 21:3–20.doi:10.5194/angeo-21–3–2003

Pinardi N, Arneri E, Crise A, Ravaioli M, Zavatarelli M. 2006.The physical, sedimentary and ecological structure andvariability of shelf areas in the Mediterranean Sea. InRobinson AR, Brink K. editors. The sea. Cambridge, USA:Harvard University Press; 14, p. 1243–1330.

Pinardi N, Coppini G. 2010. Operational oceanography in theMediterranean sea: the second stage of development. OceanSci. 6:263–267.

Pinardi N, Masetti E. 2000. Variability of the large-scale generalcirculation of the, Mediterranean sea from observations andmodelling: a review. Palaeogeography, Palaeoclimatology,Palaeoecology. 158:153–173.

Pinardi N, Zavatarelli M, Adani M, Coppini G, Fratianni C, OddoP, Simoncelli S, Tonani M, Lyubartsev V, Dobricic S, BonaduceA. 2015. Mediterranean sea large-scale low-frequency oceanvariability and water mass formation rates from 1987 to2007: a retrospective analysis. Prog Oceanogr. 132:318–332.

Poulain P-M, Barbanti R, Font J, Cruzado A, Millot C,Gertman I, Griffa A, Molcard A, Rupolo V, Le Bras S,Petit de la Villeon L. 2007. MedArgo: a drifting profiler pro-gram in the Mediterranean sea. Ocean Sci. 3:379–395.

Rinaldi E, Orasi A, Morucci S, Colella S, Inghilesi R, BignamiF, Santoleri R. (this issue). How can operational oceanogra-phy products contribute to the European Marine StrategyFramework Directive? The Italian case.

Shih LH, Koseff JR, Ivey GN, Ferziger JH. 2005.Parametrization of turbulent fluxes and scales using homo-geneous sheared stably stratified turbolence simulations. JFluid Mech. 525:193–214. doi:10.1017/S0022112004002587

JOURNAL OF OPERATIONAL OCEANOGRAPHY s233